Secondary attachment and mental health in Pakistani and Scottish adolescents: A moderated mediation model

OBJECTIVES
Research into adolescent mental health has tended to focus on primary attachment relationships. However, the effect of secondary attachment relationships and the role of culture remain under-explored. This study examined the associations between primary attachment, secondary attachment, and coping strategies (task-focused, emotion-focused, and avoidant coping) with psychological well-being and psychological distress in adolescents across two cultural settings.


DESIGN
A cross-sectional study.


METHOD
An identical test battery was used across two geographic sites in Pakistan (N = 510; 12-18 years; 51.5% male; mean age = 14.50) and Scotland (N = 610; 12-18 years; 53.6% male; mean age = 13.97). Associations were tested separately in each sample using moderated mediation modelling for the outcome variables: psychological well-being and psychological distress.


RESULTS
For psychological well-being, all three coping strategies were significant partial mediators and secondary attachment was a significant moderator in both samples. Secondary attachment moderated the association between emotion-focused coping and psychological well-being in the Pakistani sample only. For psychological distress, task-focused coping was a significant full mediator in the Pakistani sample only. In contrast, for the Scottish sample, task-focused coping and emotion-focused coping were significant partial mediators. Secondary attachment's direct effect on psychological distress was significant in both samples. Secondary attachment also moderated the association between emotion-focused coping and psychological distress in the Pakistani sample only.


CONCLUSIONS
The cross-cultural evidence suggests that alongside primary attachment, it is important to target secondary attachment through coping strategies, in order to enhance psychological well-being and lessen psychological distress in adolescents.


PRACTITIONER POINTS
Secondary attachment plays a different role from primary attachment in adolescents. Therefore, it is important to target both primary attachment and secondary attachment security to enhance psychological well-being and lessen psychological distress. Cross-cultural differences in coping suggest that differential strategies to target different coping dimensions may enhance adolescent well-being across cultures. These cross-cultural differences highlight the ethical importance of cultural sensitivity among clinicians working with adolescents globally.


Practitioner points
Secondary attachment plays a different role from primary attachment in adolescents. Therefore, it is important to target both primary attachment and secondary attachment security to enhance psychological well-being and lessen psychological distress. Cross-cultural differences in coping suggest that differential strategies to target different coping dimensions may enhance adolescent well-being across cultures. These cross-cultural differences highlight the ethical importance of cultural sensitivity among clinicians working with adolescents globally.
Adolescent well-being is recognized as a key global mental health challenge (Camilletti, 2018). Meta-analytic data from 27 countries estimated that the prevalence of mental disorders in adolescents is 13.4% (Polanczyk, Salum, Sugaya, Caye, & Rohde, 2015), with depressive and anxiety symptoms comprising the most common indicators of adolescent psychological distress (Cohen, Andrews, Davis, & Rudolph, 2018). Accordingly, research into protective factors, including attachment, coping, and well-being, can help inform our understanding of these mental health challenges (Camilletti, 2018). Furthermore, cultural milieu represents an important factor influencing these protective factors (Tang, Tse, Davidson, & Cheng, 2018). For instance, the emphasis on autonomy in individualistic cultures encourages a greater focus on individual goals (Yeh, Arora, & Wu, 2006), while collectivist cultures emphasize interdependence to promote greater connectedness with others (Albert, Trommsdorff, & Mishra, 2007). This implies that adolescent development and mental health could be influenced by the impact of culture on these factors (Albert et al., 2007). Furthermore, as global approaches to mental health promotion become more entrenched, scrutiny of cultural differences is critical to develop preventative and intervention strategies that could be adapted for individuals from different cultural backgrounds (Malti, Noam, Beelmann, & Sommer, 2016).
To date, despite the fact that the majority of the population in low-to middleincome countries (LMICs) such as South Asia are adolescents, there is a paucity of reliable data for this developmental period from these countries (The Global Burden of Disease, 2017). Importantly, while adolescents make up 38% of its total population, research on Pakistani adolescent development and mental health has only slowly emerged since the mid-1990s (Qidwai, Ishaque, Shah, & Rahim, 2010). Only limited research has simultaneously examined two independent samples of adolescents from high-income versus LMICs such as Pakistan (Khalid, Qadir, Chan, & Schwannauer, 2019). Notably, there is little research exploring protective and resilience factors in resource-restricted settings (Kieling & Rohde, 2012). Therefore, studies with meaningful representation from both high-income versus LMICs are much needed in order to proffer global mental health recommendations (Camilletti, 2018). The samples in this study from Pakistan and Scotland not only differ in terms of cultural dimensions of collectivism versus individualism (Shah & Amjad, 2011), but also represent higher versus LMICs (The World Bank List of Economics, 2018).
The current study examined direct and indirect associations between primary attachment, secondary attachment, coping, psychological well-being (autonomy, competence and relatedness), and psychological distress (depression and anxiety). This was guided by three theoretical frameworks: attachment theory, self-determination theory (SDT), and cognitive-motivational-relational theory.
Theoretical framework Attachment theory Attachment theory provides a framework for understanding the role caregiving figures play as buffers against psychological distress, with implications for enhancing psychological well-being across cultures (Li, Delvecchio, Miconi, Salcuni, & Di-Riso, 2014). Security within attachment relationships is regulated by patterns of individual helpseeking, and the attachment figures response to perceived attachment needs (Bartholomew & Horowitz, 1991). From a developmental perspective, attachment to parents (primary attachment) is postulated to form the basis for the child's development of beliefs about self and others (in terms of trust and worthiness of love and care), which in turn influence secondary attachment relationships in subsequent life (Ainsworth, 1989). Attachment research also indicates that during adolescence, individuals are more likely to direct their attachment needs towards secondary attachment figures, suggesting a shift in focus from primary to secondary attachments (Aikins, Howes, & Hamilton, 2009). Furthermore, the concept of 'earned-security' implies that individuals with insecure primary attachments may develop attachment security over time with the development of secure secondary attachment relationships (Moller, McCarthy, & Fouladi, 2002), itself suggesting a compensatory role for secondary attachment.
Research has indicated that the quality of attachment security varies across adolescence and that primary attachment is not necessarily related to later attachment (Koepke & Denissen, 2012). However, the unique role of secondary attachment in influencing mental health and well-being of adolescents has been under-explored. In particular, research exploring attachment with potential secondary attachment figures has generally considered a limited number of relationships such as media-figures (Giles & Maltby, 2004;Greene, Adams-Price, & Greene, 1990;Roberts, 2007), 'idols' (Cheung & Yue, 2012), secondary school advisors ( Van-Ryzin, Carlson, & Sroufe, 2011), and teachers (Rosmidar, 2017). Given that secondary attachment preference may differ from individual to individual, it is therefore as yet unclear whether particular categories of relationship constitute meaningful secondary attachment figure/s for individual participants. Nor can any definitive conclusions can be drawn regarding the link between secondary attachment and mental health.

Self-determination theory
The contribution of secondary attachment security towards adolescents' well-being can also be explained via SDT . SDT proposes that psychological wellbeing in adolescence can be predicted by the degree to which three basic psychological needs are met: autonomy, competence, and relatedness (Adie, Duda, & Ntoumanis, 2008;Deci et al., 2006;Patrick, Knee, Lonsbary, & Canevello, 2007). If these needs are fulfilled with reference to a particular attachment figure, this may improve psychological well-being, suggesting that adolescents may be self-motivated to shift attachment security from primary attachment to those secondary attachment figures who can satisfy needs of autonomy, competence, and relatedness (La-Guardia, Ryan, Couchman, & Deci, 2000).

Cognitive-motivational-relational theory
Cognitive-motivational-relational theory (CMRT: Folkman & Lazarus, 1985) provides a further theoretical underpinning for the role of secondary attachment relationships.
CMRT highlights the importance of motivation, emotions, and the assessment of available resources in relation to the developmental stage of adolescents (Compas, Connor-Smith, Saltzman, & Wadsworth, 2001). In adolescence, individual's 'coping repertoires' are determined by the degree to which psychological needs for autonomy, competence, and relatedness are met and are incidental to positive emotions and well-being (Aldwin, 2000). Thus, CMRT positions the shift towards secondary attachment figures as a coping mechanism to meet psychological needs (Compas et al., 2001).
Finally, coping is also relevant to adolescent secondary attachment. Herein, coping refers to cognitive and behavioural responses that aim to minimize, control, or encounter demands or challenges from internal or/and external environments that exceed individuals' resources (Folkman & Lazarus, 1980), which themselves may mediate between attachment and distress (Glazebrook, Townsend, & Sayal, 2016). In particular, caregiver's responsiveness and sensitivity at times of distress help children to develop a sense of self-capacity in their own attempts to satisfy safety and comfort needs. Securely attached individuals are more likely to engage in adaptive coping, which is itself associated with enhanced well-being (Chakroborty & Banerjee, 2017;Shaver & Mikulincer, 2004). In contrast, individuals with insecure attachment are more likely to perceive the world as unpredictable or threatening, predicting less exploration, less perceived competence, and greater use of maladaptive coping (Wei, Heppner, & Mallinckrodt, 2003). This, in turn, could heighten experiencing more psychological distress entrenching lower or less adaptive psychological well-being (Guo, Tomson, Keller, & S€ oderqvist, 2018). Individuals may use task-focused coping, (actively focusing on the stressor/s in order to solve them) or emotion-focused coping, which is focusing on the emotional distress caused by the stressors (Lazarus, 1999). Avoidant coping can be used which includes social diversion or distraction (Lazarus & Folkman, 1984). Although task-focused coping actions are associated with positive mental health outcomes (Compas, Connor-Smith, & Jaser, 2004), there have also been contradictory results regarding the effects of emotion-focused and avoidant coping on psychological well-being (Yeh et al., 2006), with culture taking an important mediating factor. Strategies such as problem solving are likely to be associated with individualistic societies while strategies of adaptation to the situation are encouraged in collectivist societies (Matsumoto, Yoo, & Nakagawa, 2008). This may suggest that cultural values such as individualism versus collectivism could affect coping.

Aims
The aim of the current study was to investigate coping as a potential mediator of the link between primary attachment and psychological distress (depression and anxiety) and psychological well-being (autonomy, competence and relatedness), and to determine whether this mediation is moderated by secondary attachment. For moderation, two conditional effects were hypothesized: (1) the direct effect of primary attachment on both outcomes: psychological well-being and psychological distress will be moderated by secondary attachment and (2) the second stage moderated mediation, which implies that secondary attachment will moderate the effects of mediators (task-focused coping, emotion-focused coping, and avoidant coping) on both outcomes ( Figure 1). Three coping subscales were considered as separate parallel mediators to examine their unique contribution to any mediation effects. It was expected that these links would vary across cultures.

Method
Design -Cross-sectional study, using an identical test battery across two geographic sites in Pakistan and Scotland.
Demographic characteristics of both samples are presented in Table 1.

Measures
For the Pakistani sample, the Urdu translated version of Hospital Anxiety and Depression Scale by Mumford, Tareen, Bajwa, and Bhatti (1991) was used. Other measures, Experiences in Close Relationship-12, Three subscales (Autonomy, Competence, and Relatedness) of Ryff's Psychological Well-being Scale and Coping Inventory for Stressful Situations-21, were back translated into Urdu (Brislin, 1976) and validated with 400 adolescents in a separate study (available on request). A demographic questionnaire was included to measure gender, age, school grade, ethnicity, relationship status, religion, and family system. The Family Affluence Scale (FAS II; Currie, 2004; a .57; Boyce, Torsheim, Currie, & Zambon, 2006) was included to assess adolescents' socio-economic status. In terms of the composite FAS II score, a score of 0-3 indicates low affluence, a score of 4-6 indicates middle affluence, and a score of 7-9 represents high affluence.
The Experiences in Close Relationship-12 (ECR-12; Wei, Russell, Mallinckrodt, & Vogel, 2007) ECR-12 was used twice (with different instructions) to measure primary attachment and secondary attachment, respectively. Therefore, this study gave the option to participants to indicate whom they considered as their primary and secondary attachment figure/s. This methodological facet enabled us to explore the independent contribution of primary and secondary attachment in adolescents' mental health. Wei et al. (2007) developed this 12-item scale from Brennan, Clark, and Shaver (1998) 36-item ECR. The ECR-12 has two subscales: anxiety and avoidance. Participants rate each question on a 7-point Likert scale.  Higher scores on both subscales were used to measure insecure attachment orientation. Wei et al. (2007) found Cronbach's alphas .78 for anxiety and .84 for avoidance in their US student sample.
The Hospital Anxiety and Depression Scale (HADS; Zigmond & Snaith, 1983) A composite score of HADS was used to measure psychological distress in terms of depression and anxiety symptoms. This is a 14-item scale with two subscales: depression and anxiety (7 items each). Participants were asked to rate each item on a 4-point Likert scale based on how they feel generally. Cronbach's a for depression and anxiety subscales was reported to be .83 and .82, respectively (Bjelland, Dahl, Haug, & Neckelmann, 2002). It has been validated for adolescents (White, Leach, Sims, Atkinson, & Cottrell, 1999).

Procedure
Prior to the main data collection, a small pilot study was conducted with 15 adolescents to estimate the time duration taken to complete the questionnaires and any difficulties in understanding instructions while answering the questions. For the main parallel surveys, a letter/email outlining the broad aim and scope was sent out to the school heads in Rawalpindi and Edinburgh. Then, a detailed information sheet and parental opt-out consent forms were sent to parents of all adolescent students of the participating schools for both samples. On the day of data collection, adolescents with a signed opt-out parental consent form were not included. For Pakistani schools, data were collected in a library hall (25-30 min) in December 2016 and April 2017. In addition to parental consent, adolescent participants were also asked to complete a consent form prior to participation. After completion, respondents were debriefed orally. In Scotland, the data were collected during study skills and religious education classes in November 2016 and May 2017, using the similar procedures to the Pakistani sample.

Ethical approval
The ethical approval received from the University of Edinburgh, Fatima Jinnah Women University Rawalpindi, Pakistan, the local education authority in Pakistan, and the City of Edinburgh Council. The participants gave consent for their data to be used in the research.

Data analysis
Following Fritz and MacKinnon (2007), a minimum sample of 405 participants was required for 0.8 power to detect an effect of 0.13 (medium effect size) with alpha 0.05. Both samples exceeded the minimum sample size of 405. Data were screened for multivariate normality, homoscedasticity, and multicollinearity. For outliers, Cook's distance, leverage, and Mahalanobis distance were examined too. Overall, no significant violation of assumptions was detected. There were no missing values as incomplete questionnaires (n = 45) were excluded before the data entry. For all chosen measures except the CISS-21, total scores were used in analyses. Descriptive statistics, reliability, and correlations (Table 2) were calculated in SPSS (Statistical Package for the Social Sciences; version 21). Moderate to high Cronbach alphas (0.71-0.94) were found across the scales in both samples in this study (see Table 3). The moderated mediation models were tested using Process Macro, Model 15 (Hayes, 2013) using 5,000 bootstrap samples. A 95% bias-corrected confidence interval was used for the index of moderated mediation (Hayes, 2015). If zero resides within the confidence interval, then the effect is nonsignificant. Therefore, the final decisions about the significance of moderated mediation were based on this. Three levels for the moderator, secondary attachment, were used based on mean and plus/minus one standard deviation (SD) from the mean. For the moderator, higher scores on ECR-12 indicated greater attachment insecurity, with the lower (1 SD below mean), middle (mean) and higher values (1 SD above mean) indicating 'high', 'medium', and 'low' secondary attachment security, respectively.

Results
No significant correlations were found between the demographic variables of age and family affluence with the outcome variables, although a number of significant differences emerged in relation to gender. Consistent with previous studies (i.e., Van Droogenbroeck, Spruyt, & Keppens, 2018;Visani et al., 2011), females reported higher psychological distress and lower psychological well-being.

Psychological well-being
In the Pakistani sample, all three mediators were significant partial mediators and secondary attachment (b = .01, SE = .004, p < .05) was the significant moderator for  psychological well-being (Figure 2). Secondary attachment also moderated (b = À.05, SE = .009, p < .01) the link between emotion-focused coping and psychological wellbeing. The effect of primary attachment on psychological well-being through emotionfocused coping was moderated by secondary attachment, and this effect became stronger with the levels of secondary attachment (Table 4). The conditional direct effects of primary attachment on psychological well-being for low (b = À.29, SE = .10, CI = À0.49 to À0.09) and middle levels of secondary attachment (b = À.16, SE = .07, CI = À0.30 to À0.01) were significant. Thus, the moderated mediation model for psychological wellbeing was supported.

Psychological distress
In the Pakistani sample, task-focused coping (b = À.55, SE =.11, p < .001) was the significant full mediator, while secondary attachment's direct effect on psychological distress was significant (b = .11, SE = .02, p < .001), which was non-significant for psychological well-being (Figure 4). The effect of primary attachment on psychological distress through emotion-focused coping was moderated by secondary attachment (b =.001, SE =.0004, 95% CI = 0.0005 to 0.002), and this effect became stronger with the levels of secondary attachment (see above Table 4). However, these conditional effects for task-focused coping and avoidant coping were non-significant. Thus, the moderated mediation model was supported. In the Scottish sample, emotion-focused coping (b = .45, SE = .04, p> .001) and taskfocused coping (b = À.22, SE = .05, p < .001) were significant partial mediators while avoidant coping's effect (b = À.07, SE = .05, p > .05) was non-significant on psychological distress ( Figure 5). Similar to the Pakistani sample, secondary attachment's direct effect (b = .11, SE = .02, p < .001) on psychological distress was significant but other conditional effects were non-significant. Overall, the moderated mediation model was partially supported.

Discussion
The current study used a theoretically driven moderated mediation model to examine associations between primary attachment, secondary attachment, and coping strategies with psychological well-being and psychological distress in Pakistani and Scottish adolescent samples. This is the first cross-cultural study in which participants were asked to identify secondary attachment. The majority of Pakistani adolescents identified siblings while Scottish adolescents identified friends as the most significant secondary attachment figure (see Table 1). These findings suggest that secondary attachment preferences may vary across individuals. Findings also further highlight the importance of letting participants self-select the secondary attachment figure/s they consider significant rather than limiting secondary attachment figures a priori.

Psychological well-being Mediation effects
In both samples, primary attachment influenced psychological well-being both directly and indirectly via adaptive coping. This is consistent with predictions from attachment theory that individuals derive adaptive coping resources from secure attachments, which unfold via consistent and responsive experiences with caregivers (Zimmer-Gembeck, Ducat, Clear, Mastro, & Van-Petegem, 2018). Positive attachment experiences with caregivers enhance individuals perceived competence and representations of others as reliable (Mallinckrodt, 2000;Wei et al., 2003). In our study, consistent with prior studies that used different indicators of psychological well-being (Abubakar, Alonso-Arbiol, Van de Vijver, Murugami, & Arasa, 2013; Ger onimo, Schoeps, & Montoya-Castilla, 2019), positive attachment experiences were associated with positive mental health outcomes. To emphasize the cultural aspects of these findings, a recent study of Chinese adolescents (Guo et al., 2018) indicated that both maternal and paternal attachments were linked to psychological resilience, both directly and indirectly through enhanced task-focused coping and reduced emotion-focused coping. This finding is also consistent with prior studies from Western cultures (Lopez, Mathers, Ezzati, Jamison, & Murray, 2006;Wei et al., 2003), reinforcing the universal role of attachment security in enhancing psychological well-being in adolescents, albeit via adaptive coping.
Moderation effects. With regard to moderation effects, for the first conditional effect, secondary attachment moderated the relationship between primary attachment and psychological well-being in both samples. As such, the negative effect of primary attachment on psychological well-being decreases with higher levels of secondary attachment security. This suggests that although insecure primary attachment contributes to reduced psychological well-being, secondary attachment security may buffer the negative impact of primary attachment insecurity on adolescents' psychological wellbeing. This is congruent with theoretical predictions and prior research on the compensatory role of secondary attachment (Shirvanian & Michael, 2017). Our finding also supports the concept of earned security (Moller et al., 2002) that secure secondary attachment buffers the effects of insecure primary attachment. The similar findings for secondary attachment across both Pakistani and Scottish adolescent samples highlight the important role of secondary attachment across cultures, particularly at this critical developmental stage when individuals are expanding their interpersonal relational networks beyond primary attachments (Aikins et al., 2009;Allen, 2008).
For the second conditional effect (second stage moderated mediation), secondary attachment moderated the relationship between emotion-focused coping and psychological well-being, albeit only in the Pakistani sample. This suggests that primary attachment insecurity leads to more emotion-focused coping, which in turns negatively affects psychological well-being but that this effect is dependent on levels of secondary attachment. Therefore, at least in the Pakistani context, it could be that secondary attachment security reduces the negative impact of insecure primary attachment through the mechanism of emotion-focused coping. Prior research indicates that a lack of emotional support from attachment figures predicts higher depression and anxiety symptoms (Pereira, Barros, Mendonc ßa, & Muris, 2014;Van Roekel, Engels, Verhagen, Goossens, & Scholte, 2011). Correspondingly, the present finding in the Pakistani sample supports a theoretical claim from CMRT that turning towards secondary attachment can be a subsequent coping strategy after cognitive appraisal of the available emotional and social support in adolescence (Aldwin, 2007;La-Guardia & Patrick, 2008). This suggests that enhancing adolescents' secondary attachment security may reduce the negative effect of emotion-focused coping on psychological well-being. The absence of this mechanism in the Scottish sample points to the cultural differences in secondary attachment relationships. Further, it has important intervention implications, as targeting coping strategies for enhancing psychological well-being via secondary attachment relationships may be more feasible in the context of youth mental health, as compared to seeking to change the quality of primary attachment relationship (Bowlby, 1982;Carr, Colthurst, Coyle, & Elliott, 2013). Finally, the current study examined psychological wellbeing in terms of autonomy, competence, and relatedness, thus consistent with the SDT; these results support the unique contribution of secondary attachment to meet basic psychological needs in adolescence (Carr et al., 2013;Ntoumanis, Edmunds, & Duda, 2009).

Psychological distress Mediation effects
In the Pakistani sample, reduced task-focused coping fully mediated the effect of primary attachment insecurity on psychological distress. This indicated that task-focused coping plays an important role in the relationship between insecure primary attachment with psychological distress. These findings are comparable to Indian findings (Shaheen & Alam, 2010), whereby the association of task-focused coping with psychological distress was significant, but its relationship with emotion-focused coping was non-significant. In traditionally collectivist countries such as India and Pakistan, emotion-focused strategies such as spending some time in seeking and giving emotional support before finding a solution to the problem are widely practised. Consequently, there is the possibility that in these cultures emotion-focused coping is not linked to psychological distress, unlike in individualistic cultural settings. Together, these results are aligned with studies across cultures (Compas et al., 2004;Guo et al., 2018) in terms of highlighting the importance of developing interventions targeting task-focused coping strategies in adolescents.
By contrast, in the Scottish sample, task-focused coping and emotion-focused coping partially mediated the relationship between primary attachment and psychological distress; the effect of avoidant coping was non-significant. This suggests that, in the Scottish sample, the effect of primary attachment insecurity on psychological distress is more likely through enhanced emotion-focused coping and reduced task-focused coping and supports the dichotomy used in coping research (Holt, Hoar, & Fraser, 2005). As such, differences in coping across Pakistani and Scottish samples support a cross-cultural approach to differences in coping (Glazer, 2006;Matsumoto et al., 2008).

Moderation effects
Regarding the first conditional effect, secondary attachment did not moderate the relationship between primary attachment and psychological distress; instead, secondary attachment's direct effect on psychological distress was significant across both samples. In contrast, for psychological well-being, secondary attachment moderated the direct link between primary attachment and psychological well-being. This suggests that insecure secondary attachment may constitute a risk factor for psychological distress in adolescents. Like the previous results, these findings are also consistent across two independent cultural samples. Nevertheless, future research should consider other indicators of psychological well-being and psychological distress to explore these associations across cultures.
For the second conditional effect (second stage moderated mediation), secondary attachment moderated the path between emotion-focused coping and psychological distress, but only in the Pakistani sample. Although the relationship between emotionfocused and psychological distress was not significant, its interaction effect with secondary attachment was significant on psychological distress. These results demonstrate that both emotion-focused coping and secondary attachment insecurity contribute to adolescent psychological distress. One explanation for this finding could be that attachment with parents (primary attachment) is likely to be more obedience-oriented while with siblings or friends (secondary attachment) more emotionally close relationship is expected within a Pakistani context (Khalid, Qadir, Chan, & Schwannauer, 2017). This suggests that targeting adolescent secondary attachment security could reduce the influence of emotion-focused coping on psychological distress.
Considered together, the current study's results highlight a network of unique links between primary attachment and secondary attachment with different coping dimensions, which in turn are associated with psychological well-being and psychological distress in adolescents across cultures. Therefore, it emphasizes that different strategies to target different coping dimensions are required for enhancing adolescent well-being across cultures. Furthermore, overall analyses from the current study showed differential patterns of relationships between our psychological mechanisms on psychological wellbeing and psychological distress. These findings are consistent with a dual continuum global mental health approach whereby mental health and well-being represent two distinct but correlated dimensions. Consequently, considering both as outcome variables will help policy makers and professionals to improve targeting of well-being enhancing strategies in adolescents (Guo et al., 2018).

Limitations
As the current study used a cross-sectional design, future longitudinal studies will be needed to establish causal links between these variables. Data were collected from an urban community population, and as such, findings may not be directly applicable to rural/clinical samples. The use of self-report measures may have introduced response bias. The measures used were screening tools; diagnostic measures are suggested for future research. For assessment of young people internalizing symptoms such as depression and anxiety, self-reports are considered more valid. Future studies may employ multiple informant approaches such as parent, teachers, or siblings' reports (Holmbeck, Li, Schurman, Friedman, & Coakley, 2002).

Conclusions
This study provides meaningful data on adolescents from both high-income and lowermiddle-income settings as well as individualistic versus collectivist societies. First, it helps advance our theoretical understanding of how primary attachment and secondary attachment are linked and to identify their independent and combined role in adolescent mental health. It supports the position of secondary attachment as unique from primary attachment. Therefore, alongside primary attachment, it is important to target secondary attachment security to enhance psychological well-being and lessen psychological distress. The theoretical foundation of our study (attachment theory, SDT, and CMRT) provides a novel framework for understanding the unique direct and indirect links between attachment relationships and coping with mental health outcomes. This has important implications for intervention and treatment for diverse cultural settings. The cross-cultural differences also highlight the ethical importance of cultural sensitivity among clinicians contributing to global approaches to adolescent mental health promotion.
Secondary attachment and mental health 15